- Email: [email protected]
The clinical importance of axillary lymphadenopathy detected on screening mammography
Clinical Radiology (1997) 52, 458-461
The Clinical Importance of Axillary Lymphadenopathy Detected Screening Mammography
on
M. E. MURRAY and R. M. GIVEN-WILSON
Department of Radiology, St George's Healthcare Trust, London, UK The aim of this study was to determine the incidence and cause of axillary lymphadenopathy detected by screening m a m m o g r a p h y and to devise a management protocol for this pathology. In a retrospective study of 95 806 consecutive screening m a m m o g r a m s , 37 cases of 'pathological' axillary nodes were identified using two or more of the following criteria: size >2 cm, replacement of fatty hilum, rounded shape and generalized increased density. In 16 cases with an additional mammographic abnormality, 12 had a mass (10 malignant and two benign) and four had suspicious calcification (all malignant). In 12 of these cases, the lymph nodes showed malignancy (75 %). In 21 patients with lymphadenopathy alone on screening, six patients had a known underlying diagnosis and were not recalled from screening. The remaining 15 patients were recalled for further assessment including fine needle aspiration cytology (FNAC). The ultimate diagnosis was benign in 10 cases (48%) six reactive changes, one healed granulomatous disease, one rheumatoid arthritis, one amyloid and one acute infection - and malignant in 11 cases (52%) - six non-Hodgkin's lymphoma, four metastatic carcinoma and one leukaemia. In conclusion, there is a high incidence of malignant nodal involvement in cases of screen detected lymphadenopathy (62% of cases in our series). We would advise that patients with lymphadenopathy as the sole finding on screening mammography and in w h o m there is no k n o w n underlying cause should undergo FNAC followed by excision biopsy. Fifty per cent of such patients in this study had underlying malignancy. Murray, M.E. & Given-Wilson, R.M. (1997). Clinical Radiology 52, 458-461. The Clinical Importance of Axillary Lymphadenopathy Detected on Screening Mammography
Accepted for Publication 22 October 1996
Medio-lateral oblique mammography, taken in all patients attending the breast screening programme, routinely demonstrates the low axilla. Normal lymph nodes are commonly demonstrated as well-defined, soft tissue density nodules that are elliptical or bean shaped and often have a lucent notch or centre representing the fatty hilum. They are usually less than 1.5 cm in diameter [1]. Occasionally, 'pathological' nodes will be demonstrated (Fig_ 1). There is no consensus on the size which constitutes a pathological node, but figures of 2-2.5 cm have been suggested [2]. Size alone is not a reliable criterion as fat replaced normal nodes measuring up to 3 cm may be seen. Other features of pathological nodes include replacement of the fatty hilum, rounded shape and generalized increased density. On mammography, lymphadenopathy may be the sole finding or may be accompanied by another abnormality such as a mass or calcifications suggesting a malignant aetiology. The aim of this study was to determine the incidence and cause of pathological lymph nodes detected on routine screening mammography and to determine a management protocol for those patients with unexplained lymphadenopathy and otherwise normal mammograms. METHODS
December 1995. All cases with pathological axillary lymph nodes detected on initial screening were identified. Where records from the screening centre or from the patient's GP indicated that the lymphadenopathy was due to a known condition which was already being managed elsewhere, she was not recalled for further assessment. When lymphadenopathy was an unexpected finding, with or without another mammographic abnormality, patients were recalled for further assessment in the screening unit. This included a full history and clinical examination, ultrasound, further mammography and FNAC as appropriate. FNAC was taken of the most accessible node under palpation or ultrasound guidance with a 21 gauge needle following infiltration of the skin and subcutaneous tissues with lignocaine. The aspirate was immediately resuspended in Hank's solution. From this cytospin preparations were made and stained with Papanicolaou and May-Grunwald-Giemsa stains. Part of the cytospin sample was reserved for immunophenotyping. When malignancy was suspected on the basis of these investigations, or i f the nodes persisted (despite benign cytology), referral was made to a surgeon for excision biopsy. We have reviewed the findings on examination and imaging in all these patients. Where appropriate, we have also reviewed cytology and histopathology results.
A retrospective review was conducted of the records of 95 806 women screened at the South West London Breast Screening Service (SWLBSS) between October 1991 and
RESULTS
Correspondence to: Dr M.E. Murray, Robinswood, 20 Park Road, Haslernere, Surrey GU27 2NL, UK.
Thirty-seven cases of pathological axillary lymphadenopathy were identified (0.039% women screened). The age
© 1997 The Royal College of Radiologists.
459
SCREEN-DETECTEDAXILLARY LYMPHADENOPATHY
Table 2 - Final diagnosis in women with lymphadenopathy as the sole mammographic abnormality (n = 21)*
Diagnosis
Underlying cause known. No recall (n = 6)
Healed granulomatous disease (n = 1) Rheumatoid arthritis (n = 1) Leukaemia (n = l) Lymphoma (n = 1)
Recalled for full assessment (n = 15)
Acute breast infection resolved on antibiotics (n = 1) Persisting nodes (n = 14) Benign cytology (C2) (n = 8) Refused surgery. Well on follow-up
(n = 1) Amyloid (n = 1) Reactive nodes (n = 1) Undifferentiated carcinoma (n = 1) Malignant cytology (C5) (n = 3) IDC + nodal metastases (n = 1) Pulmonary + nodal metastases (n = 1) Nodal metastases (n = 1) NHL on cytology (n = 3) Total: 11 malignant, l0 benign. Those assessed: 7 malignant (all with nodal metastases) and eight benign. * IDC, invasive ductal carcinoma; NHL, non-Hodgkin's lymphoma.
Fig. 1 - Pathological nodes. Large (3cm), round, dense nodes with replacement of the fatty hilum in a patient with non-Hodgkin' s lymphoma_
range of patients was from 50 to 67 (mean 62). During this period, in the population studied, 587 invasive carcinomas were detected on screening (0.61% of women screened). Of the 587 patients with invasive carcinoma, 304 (52%) underwent axillary surgery and 72 (12.3%) were found to have nodal metastases. This is likely to be an underestimate of nodal involvement, since there may be further women with nodal metastases who did not have initial axillary surgery. In 16 cases, there was an additional mammographic abnormality (see Table 1). All these women were recalled for full assessment then referred for surgical management. Fourteen patients were found to have malignant disease, 12 Table 1 - Final diagnosis in women with lymphadenopathy plus another mammographic lesion*
Breast mass (n = 12)
n
Suspicious breast calcifications (n = 4)
n
IDC + nodal metastases IDC + reactive nodes Follicular lymphoma Recurrent phylloides + reactive nodes Radial scar + BBD + reactive nodes
8 1 1
IDC -]- nodal metastases DCIS + reactive nodes
3 1
1 1
Total: 14 malignant (12 with nodal metastases) and two benign. * IDC, invasive ductal carcinoma; DCIS, ductal carcinoma in situ; and BBD, benign breast disease. © 1997 The Royal College of Radiologists, Clinical Radiology, 52, 458-461.
with nodal metastases. Two patients had benign disease (one phylloides tumour and one radial scar). In 21 women axillary lymphadenopathy was the sole mammographic abnormality (see Table 2). When no cause was known (n = 15), the women underwent full assessment with FNAC of the most accessible node, if appropriate. Malignant cytology (C5 or lymphoma) prompted immediate referral (n = 6). Those with benign cytology (n = 8) were reviewed at 6 weeks. Lymphadenopathy persisted in all cases and referral was then made for surgical excision biopsy to make a histological diagnosis. One patient found on recall to have an infected breast lesion was treated with antibiotics with full resolution. Three patients had malignant cytology suggestive of a breast primary with no breast lesion seen in imaging. In one case, axillary clearance was undertaken and 10/11 nodes were involved. In addition, a 26-mm invasive ductal carcinoma was identified in the axillary tail. In the second case, a chest X-ray revealed widespread pulmonary metastases and axillary clearance was therefore not undertaken. The patient was started on Tamoxifen. In the third case, axillary clearance was undertaken and 1/15 nodes contained metastatic carcinoma of the breast. Three patients were found to have previously undiagnosed non-Hodgkin's lymphoma. None of these patients had B-symptoms, although two had enlarged supraclavicular and cervical nodes in addition to the axillary nodes demonstrated at mammography. All patients with benign cytology had persistent lymphadenopathy at 6-week review and all were referred for excision biopsy although one patient refused surgery. At the time of review, one patient had developed dermal oederna on the side of the lymphadenopathy. Cytology in this patient revealed epithelioid granulomas and the possibility of tuberculous disease was considered.- However, excision biopsy of an axillary node revealed undifferentiated carcinoma. Another patient was known to have rheumatoid arthritis but was nevertheless referred for excision biopsy due to a strong family history of breast carcinoma. Histology showed a reactive node. In another
460
CLINICAL RADIOLOGY
Fig. 2 - Enlarged normal nodes with fatty infiltration leading to atrophy of surrounding lymphoid tissue and distension of the capsule.
Fig. 3 - Densely calcified nodes secondary to previous granulomatous disease.
case, mammography revealed unilateral axillary lymphadenopathy and since there was a past history of breast carcinoma, a recurrence was suspected. Clinical examination also showed bilateral inguinal lymphadenopathy and excision biopsy of an inguinal node confirmed amyloid. This patient was found to have a benign paraprotein in the blood but there were no features to suggest myeloma.
rheumatoid arthritis and less common arthritides such as psoriasis, systemic lupus erythematosus and scleroderma. Enlarged axillary nodes may also occur as part of the generalized adenopathy seen in sarcoid [6]. Malignant nodal involvement may be due to metastatic breast cancer, lymphoma, leukaemia or metastases from a distant primary. Kalisher et al. [2], found that when nodes were dense and greater than 2.0 cm in diameter, the true positive rate in predicting metastases was 85% and the false negative rate was 37%. For nodes greater than 2.5 cm, the rates were 100% and 41%, respectively. In a paper published in 1987 [3], Hunter et al. claimed that the presence of axillary lymph nodes in asymptomatic screened women did not increase the relative risk of developing subsequent breast carcinoma. They therefore concluded that 'the presence and pattern of axillary lymph nodes need not form part of the decision pathway which dictates the future management of asymptomatic screened women'. We would strongly disagree with this conclusion. In our series, had we disregarded axillary lymphadenopathy, seven patients with malignant disease would not have been recalled from the screening programme, four patients with metastatic carcinoma and three with non-Hodgkin's lymphoma. The mammograms in all of these patients showed axillary lymphadenopathy with no other abnormality. In our series, 16 patients presented with a mass or suspicious calcifications plus 'pathological' nodes and twelve of these had an invasive ductal carcinoma. Eleven
DISCUSSION If medio-lateral oblique mammograms are performed, approximately 50% of the screened population will have demonstrable axillary lymph nodes [3]. Normal nodes are small (<1.5 cm), round or oval and have a lucent notch or centre due to the fatty hilum. Fatty infiltration of nodes is common in elderly patients. The fat distends the capsule and enlarges the node, and the surrounding lymphoid tissue atrophies [4] (Fig. 2). These nodes may be up to 3 cm in diameter and benign [1]. Coarse calcification of nodes may occur in granulomatous infections [5] (Fig. 3). There are no strict criteria for the size of pathological axillary nodes, but those over 2.0-2.5 cm should be regarded as suspicious when at least one other abnormal feature is present, as mentioned previously. In all cases of mammographically detected 'pathological' lymph nodes, an underlying cause should be sought. Common benign causes of adenopathy include dermal infections (both of the hands and breasts),
© 1997 The Royal College of Radiologists, Clinical Radiology, 52, 458-461.
SCREEN-DETECTEDAXILLARY LYMPHADENOPATHY
of the 12 patients (92%) had metastatic nodal involvement. The presence of pathological nodes detected on mammography in association with a mass or suspicious calcifications therefore has important implications regarding the staging and management of the primary tumour, and should not be ignored [7]. It is also of note that one patient in whom cytology was benign, showing epithelioid granulomas, had metastatic carcinoma on histology. For this reason, it is important to follow-up patients with benign cytology after a short interval. We chose 6 weeks as an arbitrary follow-up period. However, none of the pathological nodes resolved in this time except those in the one case with clinically obvious infection. It may be preferable to proceed to immediate excision of an abnormal node if there is no clinically obvious infective aetiology. This would expedite the diagnosis in those found to have malignancy. It was interesting how low the overall sensitivity of mammography was for detecting lymphadenopathy in the total number of women in this series found to have nodal metastases at surgery. In the total screening population studied (95 806 cases), 72 women had invasive breast cancer with nodal metastases found at operation. In only 12 of these was lymphadenopathy detected on initial screening mammograms (12/72, 17% sensitivity). This may be partly due to the fact that the majority of axillary nodes are not included in the oblique mammogram. A further contributory factor is likely to be the size of the metastases. A deposit which does not enlarge the node or replace the fatty hilum will not be detected even if that node is included on the mammogram. The absence of lymphadenopathy on mammography in a patient with breast carcinoma is a poor indicator of the presence or absence of nodal metastases.
© 1997 The Royal College of Radiologists, ClinicalRadiology, 52, 458-461.
461
When lymphadenopathy is seen on screening films, the specificity for malignant nodal involvement, either breast carcinoma or haematological, is high (23/37, 62%). In patients in whom it is the sole mammographic finding, and in whom there is no known underlying cause, ultrasound examination and FNAC followed by excision biopsy should be carried out. In this series, 50% of such women had underlying malignancy. In patients with a suspicious mass or calcifications plus 'pathological' nodes, the surgeon should be alerted to the high probability of nodal metastases (92% of patients with IDC in our series).
Acknowledgement. The authors wish to acknowledge the assistance of Ms Janet Ansell, Deputy QA Manager, South Thames West Quality Assurance Reference Centre.
REFERENCES 1 Kalisher L. Xeroradiography of axillary lymph node disease_ Radiology 1975;114:67-71. 2 Kalisher L, Chu AM, Peyster RG. Clinicopathological correlation of xeroradiography in determining involvement of metastatic axillary nodes in female breast cancer. Radiology 1976;121:333-335. 3 Hunter JV, Hunter GJ, Tucker AK. Patterns of axillary lymphadenopathy demonstrated by mammography: implications for the asymptomatic women in a breast screening programme. Clinical Radiology 1987; 38:515-517. 4 Leborgne R, Leborgne F, Leborgne JH. Soft-tissue radiography of axillary nodes with fatty infiltration. Radiology 1965;84:513-515. 5 Shaw de Paredes E. Atlas of Film-screen Mammography, 2nd ed. Baltimore: Williams & Wilkins, 1991. 6 Lazarus AA. Sarcoidosis. Otolaryngology clinics of North America 1982;15(3):621-633. 7 Rayter Z, Williams J, Gazet J-C. The significance of lymph nodes detected by xeroradiography in cancer of the breast. Clinical Oncology 1981;7:39-43.
The Clinical Importance of Axillary Lymphadenopathy Detected Screening Mammography
on
M. E. MURRAY and R. M. GIVEN-WILSON
Department of Radiology, St George's Healthcare Trust, London, UK The aim of this study was to determine the incidence and cause of axillary lymphadenopathy detected by screening m a m m o g r a p h y and to devise a management protocol for this pathology. In a retrospective study of 95 806 consecutive screening m a m m o g r a m s , 37 cases of 'pathological' axillary nodes were identified using two or more of the following criteria: size >2 cm, replacement of fatty hilum, rounded shape and generalized increased density. In 16 cases with an additional mammographic abnormality, 12 had a mass (10 malignant and two benign) and four had suspicious calcification (all malignant). In 12 of these cases, the lymph nodes showed malignancy (75 %). In 21 patients with lymphadenopathy alone on screening, six patients had a known underlying diagnosis and were not recalled from screening. The remaining 15 patients were recalled for further assessment including fine needle aspiration cytology (FNAC). The ultimate diagnosis was benign in 10 cases (48%) six reactive changes, one healed granulomatous disease, one rheumatoid arthritis, one amyloid and one acute infection - and malignant in 11 cases (52%) - six non-Hodgkin's lymphoma, four metastatic carcinoma and one leukaemia. In conclusion, there is a high incidence of malignant nodal involvement in cases of screen detected lymphadenopathy (62% of cases in our series). We would advise that patients with lymphadenopathy as the sole finding on screening mammography and in w h o m there is no k n o w n underlying cause should undergo FNAC followed by excision biopsy. Fifty per cent of such patients in this study had underlying malignancy. Murray, M.E. & Given-Wilson, R.M. (1997). Clinical Radiology 52, 458-461. The Clinical Importance of Axillary Lymphadenopathy Detected on Screening Mammography
Accepted for Publication 22 October 1996
Medio-lateral oblique mammography, taken in all patients attending the breast screening programme, routinely demonstrates the low axilla. Normal lymph nodes are commonly demonstrated as well-defined, soft tissue density nodules that are elliptical or bean shaped and often have a lucent notch or centre representing the fatty hilum. They are usually less than 1.5 cm in diameter [1]. Occasionally, 'pathological' nodes will be demonstrated (Fig_ 1). There is no consensus on the size which constitutes a pathological node, but figures of 2-2.5 cm have been suggested [2]. Size alone is not a reliable criterion as fat replaced normal nodes measuring up to 3 cm may be seen. Other features of pathological nodes include replacement of the fatty hilum, rounded shape and generalized increased density. On mammography, lymphadenopathy may be the sole finding or may be accompanied by another abnormality such as a mass or calcifications suggesting a malignant aetiology. The aim of this study was to determine the incidence and cause of pathological lymph nodes detected on routine screening mammography and to determine a management protocol for those patients with unexplained lymphadenopathy and otherwise normal mammograms. METHODS
December 1995. All cases with pathological axillary lymph nodes detected on initial screening were identified. Where records from the screening centre or from the patient's GP indicated that the lymphadenopathy was due to a known condition which was already being managed elsewhere, she was not recalled for further assessment. When lymphadenopathy was an unexpected finding, with or without another mammographic abnormality, patients were recalled for further assessment in the screening unit. This included a full history and clinical examination, ultrasound, further mammography and FNAC as appropriate. FNAC was taken of the most accessible node under palpation or ultrasound guidance with a 21 gauge needle following infiltration of the skin and subcutaneous tissues with lignocaine. The aspirate was immediately resuspended in Hank's solution. From this cytospin preparations were made and stained with Papanicolaou and May-Grunwald-Giemsa stains. Part of the cytospin sample was reserved for immunophenotyping. When malignancy was suspected on the basis of these investigations, or i f the nodes persisted (despite benign cytology), referral was made to a surgeon for excision biopsy. We have reviewed the findings on examination and imaging in all these patients. Where appropriate, we have also reviewed cytology and histopathology results.
A retrospective review was conducted of the records of 95 806 women screened at the South West London Breast Screening Service (SWLBSS) between October 1991 and
RESULTS
Correspondence to: Dr M.E. Murray, Robinswood, 20 Park Road, Haslernere, Surrey GU27 2NL, UK.
Thirty-seven cases of pathological axillary lymphadenopathy were identified (0.039% women screened). The age
© 1997 The Royal College of Radiologists.
459
SCREEN-DETECTEDAXILLARY LYMPHADENOPATHY
Table 2 - Final diagnosis in women with lymphadenopathy as the sole mammographic abnormality (n = 21)*
Diagnosis
Underlying cause known. No recall (n = 6)
Healed granulomatous disease (n = 1) Rheumatoid arthritis (n = 1) Leukaemia (n = l) Lymphoma (n = 1)
Recalled for full assessment (n = 15)
Acute breast infection resolved on antibiotics (n = 1) Persisting nodes (n = 14) Benign cytology (C2) (n = 8) Refused surgery. Well on follow-up
(n = 1) Amyloid (n = 1) Reactive nodes (n = 1) Undifferentiated carcinoma (n = 1) Malignant cytology (C5) (n = 3) IDC + nodal metastases (n = 1) Pulmonary + nodal metastases (n = 1) Nodal metastases (n = 1) NHL on cytology (n = 3) Total: 11 malignant, l0 benign. Those assessed: 7 malignant (all with nodal metastases) and eight benign. * IDC, invasive ductal carcinoma; NHL, non-Hodgkin's lymphoma.
Fig. 1 - Pathological nodes. Large (3cm), round, dense nodes with replacement of the fatty hilum in a patient with non-Hodgkin' s lymphoma_
range of patients was from 50 to 67 (mean 62). During this period, in the population studied, 587 invasive carcinomas were detected on screening (0.61% of women screened). Of the 587 patients with invasive carcinoma, 304 (52%) underwent axillary surgery and 72 (12.3%) were found to have nodal metastases. This is likely to be an underestimate of nodal involvement, since there may be further women with nodal metastases who did not have initial axillary surgery. In 16 cases, there was an additional mammographic abnormality (see Table 1). All these women were recalled for full assessment then referred for surgical management. Fourteen patients were found to have malignant disease, 12 Table 1 - Final diagnosis in women with lymphadenopathy plus another mammographic lesion*
Breast mass (n = 12)
n
Suspicious breast calcifications (n = 4)
n
IDC + nodal metastases IDC + reactive nodes Follicular lymphoma Recurrent phylloides + reactive nodes Radial scar + BBD + reactive nodes
8 1 1
IDC -]- nodal metastases DCIS + reactive nodes
3 1
1 1
Total: 14 malignant (12 with nodal metastases) and two benign. * IDC, invasive ductal carcinoma; DCIS, ductal carcinoma in situ; and BBD, benign breast disease. © 1997 The Royal College of Radiologists, Clinical Radiology, 52, 458-461.
with nodal metastases. Two patients had benign disease (one phylloides tumour and one radial scar). In 21 women axillary lymphadenopathy was the sole mammographic abnormality (see Table 2). When no cause was known (n = 15), the women underwent full assessment with FNAC of the most accessible node, if appropriate. Malignant cytology (C5 or lymphoma) prompted immediate referral (n = 6). Those with benign cytology (n = 8) were reviewed at 6 weeks. Lymphadenopathy persisted in all cases and referral was then made for surgical excision biopsy to make a histological diagnosis. One patient found on recall to have an infected breast lesion was treated with antibiotics with full resolution. Three patients had malignant cytology suggestive of a breast primary with no breast lesion seen in imaging. In one case, axillary clearance was undertaken and 10/11 nodes were involved. In addition, a 26-mm invasive ductal carcinoma was identified in the axillary tail. In the second case, a chest X-ray revealed widespread pulmonary metastases and axillary clearance was therefore not undertaken. The patient was started on Tamoxifen. In the third case, axillary clearance was undertaken and 1/15 nodes contained metastatic carcinoma of the breast. Three patients were found to have previously undiagnosed non-Hodgkin's lymphoma. None of these patients had B-symptoms, although two had enlarged supraclavicular and cervical nodes in addition to the axillary nodes demonstrated at mammography. All patients with benign cytology had persistent lymphadenopathy at 6-week review and all were referred for excision biopsy although one patient refused surgery. At the time of review, one patient had developed dermal oederna on the side of the lymphadenopathy. Cytology in this patient revealed epithelioid granulomas and the possibility of tuberculous disease was considered.- However, excision biopsy of an axillary node revealed undifferentiated carcinoma. Another patient was known to have rheumatoid arthritis but was nevertheless referred for excision biopsy due to a strong family history of breast carcinoma. Histology showed a reactive node. In another
460
CLINICAL RADIOLOGY
Fig. 2 - Enlarged normal nodes with fatty infiltration leading to atrophy of surrounding lymphoid tissue and distension of the capsule.
Fig. 3 - Densely calcified nodes secondary to previous granulomatous disease.
case, mammography revealed unilateral axillary lymphadenopathy and since there was a past history of breast carcinoma, a recurrence was suspected. Clinical examination also showed bilateral inguinal lymphadenopathy and excision biopsy of an inguinal node confirmed amyloid. This patient was found to have a benign paraprotein in the blood but there were no features to suggest myeloma.
rheumatoid arthritis and less common arthritides such as psoriasis, systemic lupus erythematosus and scleroderma. Enlarged axillary nodes may also occur as part of the generalized adenopathy seen in sarcoid [6]. Malignant nodal involvement may be due to metastatic breast cancer, lymphoma, leukaemia or metastases from a distant primary. Kalisher et al. [2], found that when nodes were dense and greater than 2.0 cm in diameter, the true positive rate in predicting metastases was 85% and the false negative rate was 37%. For nodes greater than 2.5 cm, the rates were 100% and 41%, respectively. In a paper published in 1987 [3], Hunter et al. claimed that the presence of axillary lymph nodes in asymptomatic screened women did not increase the relative risk of developing subsequent breast carcinoma. They therefore concluded that 'the presence and pattern of axillary lymph nodes need not form part of the decision pathway which dictates the future management of asymptomatic screened women'. We would strongly disagree with this conclusion. In our series, had we disregarded axillary lymphadenopathy, seven patients with malignant disease would not have been recalled from the screening programme, four patients with metastatic carcinoma and three with non-Hodgkin's lymphoma. The mammograms in all of these patients showed axillary lymphadenopathy with no other abnormality. In our series, 16 patients presented with a mass or suspicious calcifications plus 'pathological' nodes and twelve of these had an invasive ductal carcinoma. Eleven
DISCUSSION If medio-lateral oblique mammograms are performed, approximately 50% of the screened population will have demonstrable axillary lymph nodes [3]. Normal nodes are small (<1.5 cm), round or oval and have a lucent notch or centre due to the fatty hilum. Fatty infiltration of nodes is common in elderly patients. The fat distends the capsule and enlarges the node, and the surrounding lymphoid tissue atrophies [4] (Fig. 2). These nodes may be up to 3 cm in diameter and benign [1]. Coarse calcification of nodes may occur in granulomatous infections [5] (Fig. 3). There are no strict criteria for the size of pathological axillary nodes, but those over 2.0-2.5 cm should be regarded as suspicious when at least one other abnormal feature is present, as mentioned previously. In all cases of mammographically detected 'pathological' lymph nodes, an underlying cause should be sought. Common benign causes of adenopathy include dermal infections (both of the hands and breasts),
© 1997 The Royal College of Radiologists, Clinical Radiology, 52, 458-461.
SCREEN-DETECTEDAXILLARY LYMPHADENOPATHY
of the 12 patients (92%) had metastatic nodal involvement. The presence of pathological nodes detected on mammography in association with a mass or suspicious calcifications therefore has important implications regarding the staging and management of the primary tumour, and should not be ignored [7]. It is also of note that one patient in whom cytology was benign, showing epithelioid granulomas, had metastatic carcinoma on histology. For this reason, it is important to follow-up patients with benign cytology after a short interval. We chose 6 weeks as an arbitrary follow-up period. However, none of the pathological nodes resolved in this time except those in the one case with clinically obvious infection. It may be preferable to proceed to immediate excision of an abnormal node if there is no clinically obvious infective aetiology. This would expedite the diagnosis in those found to have malignancy. It was interesting how low the overall sensitivity of mammography was for detecting lymphadenopathy in the total number of women in this series found to have nodal metastases at surgery. In the total screening population studied (95 806 cases), 72 women had invasive breast cancer with nodal metastases found at operation. In only 12 of these was lymphadenopathy detected on initial screening mammograms (12/72, 17% sensitivity). This may be partly due to the fact that the majority of axillary nodes are not included in the oblique mammogram. A further contributory factor is likely to be the size of the metastases. A deposit which does not enlarge the node or replace the fatty hilum will not be detected even if that node is included on the mammogram. The absence of lymphadenopathy on mammography in a patient with breast carcinoma is a poor indicator of the presence or absence of nodal metastases.
© 1997 The Royal College of Radiologists, ClinicalRadiology, 52, 458-461.
461
When lymphadenopathy is seen on screening films, the specificity for malignant nodal involvement, either breast carcinoma or haematological, is high (23/37, 62%). In patients in whom it is the sole mammographic finding, and in whom there is no known underlying cause, ultrasound examination and FNAC followed by excision biopsy should be carried out. In this series, 50% of such women had underlying malignancy. In patients with a suspicious mass or calcifications plus 'pathological' nodes, the surgeon should be alerted to the high probability of nodal metastases (92% of patients with IDC in our series).
Acknowledgement. The authors wish to acknowledge the assistance of Ms Janet Ansell, Deputy QA Manager, South Thames West Quality Assurance Reference Centre.
REFERENCES 1 Kalisher L. Xeroradiography of axillary lymph node disease_ Radiology 1975;114:67-71. 2 Kalisher L, Chu AM, Peyster RG. Clinicopathological correlation of xeroradiography in determining involvement of metastatic axillary nodes in female breast cancer. Radiology 1976;121:333-335. 3 Hunter JV, Hunter GJ, Tucker AK. Patterns of axillary lymphadenopathy demonstrated by mammography: implications for the asymptomatic women in a breast screening programme. Clinical Radiology 1987; 38:515-517. 4 Leborgne R, Leborgne F, Leborgne JH. Soft-tissue radiography of axillary nodes with fatty infiltration. Radiology 1965;84:513-515. 5 Shaw de Paredes E. Atlas of Film-screen Mammography, 2nd ed. Baltimore: Williams & Wilkins, 1991. 6 Lazarus AA. Sarcoidosis. Otolaryngology clinics of North America 1982;15(3):621-633. 7 Rayter Z, Williams J, Gazet J-C. The significance of lymph nodes detected by xeroradiography in cancer of the breast. Clinical Oncology 1981;7:39-43.